I wrote this fast implementation for PE problems as well, but mostly for
SOLSTRAS challenge at spoj.pl.
It is based on "memoryless" variant of the sieve of Eratosthenes (over priority
queue). I heard that sieve of Atkins may be much faster, but cannot investigate
it right now.
A good test may be
Thank you for assistance!
That seems to be output from an older version (before rev. 9) of the program.
How get you the code? I tried bzr branch lp:primes/ from a clean place, built
via make, run it as
./primes64 775773 775787
and it prints:
start=775773, stop=775787, mode=0x0,
Project https://launchpad.net/primes/ was designed to fix this issue. It
contains primes32, that support only 32-bit numbers, and primes64, that
supports 64-bit numbers. It is somewhat incompatible with the original primes
program, as it, if only 1 argument x given, prints prime numbers from 0 t
